Launching today
VelocityDB
Thread-per-core Rust in-memory database (1.6M+ RPS)
2 followers
Thread-per-core Rust in-memory database (1.6M+ RPS)
2 followers
VelocityDB is a memory-safe, thread-per-core database built in Rust using Linux io_uring. Designed to maximize multi-core hardware utilization, it handles 1.6M+ RPS with sub-millisecond latency. It includes a real-time Next.js observability studio.
Free
Launch Team / Built With
Render Sandboxes: a computer for every agentGive every agent its own machine. Starts in under a second
Promoted
Hi Product Hunt! 👋 I’m Yasir, the creator of VelocityDB.
I built VelocityDB because I was tired of seeing modern multi-core servers being severely underutilized by legacy, single-threaded databases—which ultimately leads to massive cloud scaling bills.
To solve this, I engineered a shared-nothing, thread-per-core database from the ground up in Rust utilizing io_uring. Instead of dealing with cross-core locking and context switching, each core manages its own memory and network I/O independently.
The Benchmark:
I recently stress-tested the engine on a basic AWS t3.medium (2 vCPUs) free-tier instance. It processed 25,000,000 requests in just 70 seconds (~357,000 RPS) with a memory footprint of under 3MB. On standard 16-core bare-metal servers, it scales linearly to over 1.6M+ RPS.
Observability Built-In:
High-performance databases shouldn't be blind black boxes. I paired the engine with a real-time Next.js Management Studio that streams hardware telemetry and RPS limits via WebSockets.
The core engine is open-sourced (AGPL v3) for the community, with Enterprise commercial licensing available for B2B deployment.
I would absolutely love for the systems engineers and founders here to tear apart the architecture, test the deployment, and share your brutal feedback! 🚀